This AI Startup, Backed By NVIDIA, Plans To Design Its Own Computer Chips

The company aims to reduce costs and gain more control over the technology powering its artificial intelligence systems.

Advertisement
Read Time: 2 mins
Mistral presently relies on Nvidia's chips.
Unsplash/ Daniel Shapiro
Quick Read
Summary is AI-generated, newsroom-reviewed
  • French AI startup Mistral plans to design its own computer chips.
  • The company currently uses Nvidia's chips.
  • The move would put Mistral alongside companies such as OpenAI, Google and Amazo
Did our AI summary help?
Let us know.

French artificial intelligence startup Mistral is planning to explore the possibility of designing its own computer chips, the company's CEO has said.

The company aims to reduce costs and gain more control over the technology powering its artificial intelligence systems. It has now raised $830 million in debt financing to fund its first dedicated data centre in Bruyeres-le-Chatel, a town near Paris in France.

Speaking to CNBC about developing its own chips, Mistral CEO Arthur Mensch said, "Of course, it is interesting. The company is not ruling it out." The move would put Mistral alongside major AI companies such as OpenAI, Google and Amazon, which are also investing heavily in custom-built chips for AI.

Currently, AI companies rely heavily on expensive chips made by firms like Nvidia. These chips are used to train AI models. With the rise in demand, the cost of running these systems has also surged.

"Owning the chips may come, I think it should come at some point, but for now we are relying on Nvidia, which is a great partner to us, and we're testing a few things here and there," he said, adding that developing custom chips could "lower the cost of deploying tokens to meaningful extents."


Mistral's facility will contain 13,800 Nvidia GB300 GPUs, which are some of Nvidia's latest advanced AI chips built specifically for heavy AI workloads. Graphics Processing Units (GPUs) are highly powerful computer chips designed to process huge amounts of data very quickly. These GPUs will deliver 44 megawatts of compute capacity.


"Europe is lagging behind when it comes to [the] buildout of infrastructure, and so we are investing to close that gap," Mensch said.


Mistral has invested around 4 billion euros in data centres in France and Sweden to expand its AI computing power. It had earlier invested about $1.4 billion in a project in Sweden. The company has also set a target of 200 megawatts of total AI compute capacity across Europe by the end of 2027


Mistral says the facility is expected to begin operations in the second quarter of 2026.

Featured Video Of The Day
Explained: How NASA Plans To Build A Base On The Moon
Topics mentioned in this article